Evidence Based Practice: finding the best evidence in 5 ...€¦ · finding the best evidence in 5 minutes Yolande Hanssens SIG Leader Medicine Information Clinical Pharmacy Coordinator

ESCP-WS SIG-MI, Lyon, October 2010 1

Evidence Based Practice: finding the best evidence

in 5 minutesYolande Hanssens

SIG Leader Medicine InformationClinical Pharmacy Coordinator & Clinical Pharmacist SICU/TICU

Hamad General Hospital – Doha - Qatar

Barbara Claus Clinical Pharmacist ICU

University Hospital – Ghent - Belgium

ESCP Lyon21-23 October 2010

2

Who are we ?

[email protected] [email protected]


3

Objectives

• To define a good clinical research question• To use some world wide (free access)

databases to search a quick answer (without obtaining more than 20 hits)

• To screen the answer for internal validity• To adopt the principle: “if the answer does

not meet the criteria of workload, validity and relevance, then the effort to explore further is not worth it”.


Outline of WS

• Introduction: EBP, RRR, ARR, NNT…• Illustrated examples• Review of evidence in small groups• Feedback to the audience• Tips & Tricks for “quick” appraisal• Summary & take home messages

IntroductionWhat is Evidence Based

Practice?Sackett DL et al, BMJ 1996; 312: 71-72

“Evidence based medicine is the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients. The practice of evidence based medicine means integrating individual clinical expertise with the best available external clinical evidencefrom systematic research.”


Hierarchy of Evidence

• Meta-analysis (MA) of several, similar, large well designed randomised controlled trials (RCTs)

• Large well designed RCT• MA of smaller RCTs• Case control and cohort studies• Case report and case series• Consensus from expert panels• I think…

The Evidence Pyramid


Sources of InformationPrimary Secondary,

Tertiary

WHAT Original researche.g. clinical trials,observational studies, case reports

Reviews, meta-analyses, interpretations, evaluations based on primary data

WHERE Scientific journals, abstracts from conferences

Special issues, editorials, clinical guidelines e-sources, etc.

TOP Pharmaceutical Journals– American Journal of Health-system

Pharmacists (US)• www.ashp.org/public/pubs/ajhp/

– The Annals of Pharmacotherapy (US)• www.theannals.com

– Pharmacotherapy (US)• www.accp.com

– European Journal of Hospital Pharmacy Practice

• www.eahp.org

– Pharmacy World and Science (EUR)• www.escpweb.org

– Clinical Therapeutics

– Clinical Pharmacy Europe

– Journal of Pharmacy Practice and Research (Australia)

• www.shpa.org.au/

– Drugs• www.adis.com/

– Drug Safety• www.adis.com/

– Préscrire• www.prescrire.org

– Journal of Clinical Pharmacy and Therapeutics

– Journal de Pharmacie Clinique

– International Journal of Pharmaceutical Compounding


TOP Medical Journals– British Medical Journal

• www.bmj.com

– Journal of the American Medical Association (JAMA)

• www.jama.ama-assn.org

– New England Journal of Medicine (NEJM)

• www.content.nejm.org

– The Lancet• www.thelancet.com

– Annals of Internal Medicine• www.annals.org

– Archives of Internal Medicine

• www.archinte.ama-assn.org

– American Journal of Medicine

• www.amjmed.org

Level of Importance

impact factor

www.isiknowledge.com/jcr

Other useful sources• Professional organizations

– Pharmacy oriented• ACCP, ASHP, ESCP etc. • Also EMEA, FDA

– Medical/Clinical oriented• General and by specialty

• Other– Medscape.com– MDLink– Etc. etc.


How/Where to begin?• 1500 pages indexed in Medline each day…• Abstracts lie (11% of abstracts contain statements that are

not present in the full article; 19% with statements inconsistent with the full article). Pitkin RM et al. JAMA 1999; 281: 1110-1.

• RCTs do not report all outcomes. Chan A-W, Altman DG. BMJ 2005; 330: 753-6.

• Doctors spent an average of less than 2 minutes pursuing an answer, and they used readily available print and human resources

• Only 2 questions (out of over 1100) led to a formal literature search. Ely JW et al. BMJ 1999; 31: 358-61

Unclear results

Douillard JY et al. Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for metastatic colorectal cancer: a multicentre randomised trial. Lancet 2000; 355: 1041-7

Time to progression


More consistent results

What to take into account?Barber N. BMJ 1995; 310: 923-5

EFFECTIVE SAFE

COST PATIENTFACTORS


What are the criteria used when looking for the “best answer”?

Slawson DC and Shaughnessy AF. J Am Board Fam Pract 1999; 12: 444-9

Usefulness = Relevance x ValidityWork

Outline of WS



Terminology: minitest

You have to select a cardiac rehabilitation program for a patient who suffered a MI, you have to decide quickly:

• Program A reduces the rate of death by 19% • Program B produces an absolute reduction in deaths of 3% • Program C increases patients survival rate from 84% to 87% • Program D means that 33 people needed to enter the program

to avoid one death

WHICH ONE WOULD YOU CHOOSE?

40

30

107.5

0

5

10

15

20

25

30

35

40

45

High risk patients Low risk patients

Eve

nt

rate

% Trial 1 CONTROL

Trial 1 ACTIVE

Trial 2 CONTROL

Trial 2 ACTIVE

Terminology: more than p-values only!


To summarize

• The RRR stays constant in different populations

• The ARR alters in different populations and will be much larger in i.e. patients with high baseline risk or in patients with a lot of events

To go back to the minitest• Program A reduces the rate of death by 19%

(RRR = 16%-13%/16%)• Program B produces an absolute reduction in deaths

of 3% ARR = 3%

• Program C increases patients survival rate from 84% to 87%16%-13% death rates = ARR = 3%

• Program D means that 33 people needed to enter the programme to avoid one death100/ARR = 33; ARR = 3%

WHICH ONE TO CHOOSE?


Outline of WS


PICO Principle• Incomplete questions challenge to

find answers in medical literature• Dissecting the question AND

restructuring the question easy to find the answers

• Most questions can be divided into 4 components

PICO


PICO PrinciplePopulation & Clinical problem

relevant people in relation to the clinical problem you have in mind

Intervention (or Indicator or Index text)shows the management strategy, exposure or test that you want to find out about in relation to the clinical problem

Procedure such as drug treatment, surgery or diet Exposure or factor that might affect a health outcome Diagnostic test, such as blood test or brain scan

Comparatorshows an alternative or control strategy, exposure or test for comparison with the one you are interested in

Outcomewhat you are most concerned about happening (or not) AND/ORwhat the patient is most concerned about

PICO Question I

Is it useful to have a clinical pharmacist in your hospital specific for antibiotics?

In other words:

Is clinical pharmacy focussing on antibiotics cost effective for your

hospital?

12/05/2016


BUILD YOUR CLINICAL QUESTION: PICO

Population = patients on antibiotic treatment

Intervention = clinical pharmacist

Comparator = regular practice/no pharmacist

Outcome = cost

(intervention question: gold standard = systematic review; RCT)12/05/2016

BUILD THE BOOLEAN EXPRESSION

TIP: use words of the PICO: antibiotic treatment, clinical pharmacist, effective

12/05/2016

antibiotic* AND (clinical pharm*) AND cost*effective AND hospital NOT ambulatory NOT primary

Wildcard (*) BOOLEANS IN CAPITALS


12/05/2016 Yolande Hanssens, Barbara Claus ESCP

LYON 2010

WWW.PUBMED.ORG


LYON 2010

antibiotic* AND (clinical pharm*) AND cost*effective ANDhospital NOT ambulatory NOT primary

2008


12/05/2016

Curr Opin Infect Dis. 2008 Aug;21(4):344‐9.

Hospital antibiotic stewardship.Lesprit P, Brun‐Buisson C.Infection Control Unit, Université Paris 12, Assistance Publique‐Hôpitaux de Paris, Hopital Henri Mondor, Créteil, France.

AbstractPURPOSE OF REVIEW: Antibiotic stewardship is designed to optimize antimicrobial therapy administered to hospitalized

patients, to ensure cost‐effective therapy and improve patients' outcome while containing bacterial resistance. Current data on the development of effective programmes, including guidelines for their implementation, have demonstrated some efficacy and controversies are reviewed.

RECENT FINDINGS: Guidelines have been recently issued for the development and implementation of active antibiotic stewardship programmes in hospitals. A multidisciplinary team including at least an infectious disease physician and a clinical pharmacist is required. Multiple strategies are available, including prospective audit with feedback to the provider, education and antimicrobial restriction. Interventions have shown a positive effect on optimization of antimicrobial use, reduced costs and bacterial

resistance, but studies showing improvement in patient outcomes are sparse. Results of studies may be

confounded by several factors, mainly due to their before‐after design and lack of control for co‐interventions.

SUMMARY: Combined with an effective infection control programme, antibiotic stewardship can help contain antimicrobial resistance. Studies demonstrating improvement of patients' outcomes are needed to increase acceptance by a broader audience of physicians. A proactive strategy of prospective auditing with direct counsels and feedback to the prescriber, ensuring systematic reassessment of ongoing therapy, appears most useful.COMMENTS:MULTIDISCIPLINARITY ...

PICO Question II At a routine immunisation visit, Lisa, the mother of asix-month-old, tells you that her baby suffered a nasty local reaction after her previous immunisation.Lisa is very concerned that the same thing may happen again this time. Recently, a colleague told you that needle length can affect local reactions to immunisation in young children but you can’t remember the exact details.

Develop a clinical research question using PICO to help you quickly find the information you

need


PICO Question IIPopulation/problem

= infants receiving immunizationIntervention

= longer needlesComparator

= shorter needles (needle length)Outcome

= local reactions

CRQ= in infants receiving immunisation injections, does needle length affect the rate of local reactions?

CRQ = Clinical Research Question

12/05/2016

WWW.PUBMED.ORG


Infants AND immunization AND needle length AND local reaction*

J Adv Nurs. 2004 Apr;46(1):66-77.A systematic review to ascertain whether the standard needle is more effective than a longer or wider needle in reducing the incidence of local reaction in children receiving primary immunization.

CONCLUSION: (1). The results are statistically very highly significant and favour the use of the 25 millimetre long needle. (2). No studies were identified to separate the effect of the gauge of the needle used, but the discussion shows that needle length is the most important variable. (3). The results should be incorporated into any future guidelines on vaccine administration in the UK.

PICO Question III Jean is a 55-year-old woman who quite often crosses the

Atlantic to visit her elderly mother. She tends to get swollen legs on these flights and is worried about her risk of developing deep vein thrombosis (DVT) because she has read quite a bit about this in the newspapers lately.

She asks you if she could wear elastic stockings on her next trip to reduce her risk of this.

How do you convert this to an answerable question, using the PICO method


PICO Question IIIPopulation/problem

= passengers on long-haul flightsIntervention

= wearing elastic compression stockingsComparator

= no elastic stockings Outcome

= development of DVT

CRQ = In passengers on long-haul flights, does wearing elastic compression stockings, compared with not wearing elastic stockings, prevent DVT?

Question Transform a clinical question into search

terminology

In passengers on long-haul flights, does wearing elastic compression stockings, compared with not wearing elastic stockings, prevent DVT?

Using BOOLEANS in Capitals

(flight OR travel) AND stocking* AND(DVT OR thrombosis)


WWW.COCHRANE.ORG

Compression stockings for preventing deep vein thrombosis in airline passengers.


2009

PICO Question IVA newly diagnosed patient, female with multiple sclerosis asks about the short term riskof becoming dependent of a wheel chair. To answer her question you perform a 1-minute search for the risk ...

Make a PICO and valid research question


BUILD YOUR CLINICAL QUESTION

12/05/2016

What is the prognosis of disability in multiple sclerosis?

PICO Question IV answertype prognosis/prediction

Population/problem= (newly diagnosed) diagnosis of multiple sclerosis

Intervention = ?

Comparator= ?

Outcome = disability


Short version P(IC)O Question IV short answer type prognosis

Population/problem= (newly diagnosed) diagnosis of multiple sclerosis

Outcome= disability


LYON 2010

WWW.PUBMED.ORG



LYON 2010

Multiple sclerosis* AND (disability OR wheel chair) AND prognos*

12/05/2016

Eur J Neurol 2010 Apr 9 [Epub ahead of print]Social consequences of multiple sclerosis: clinical and demographic predictors ‐ a historical prospective cohort study.Pfleger CC, Flachs EM, Koch‐Henriksen N.Department of Neurology, Aarhus University Hospital in Aalborg, Aalborg.

AbstractBackground: Time to disability pension is one of the endpoints to be used to determine the prognosis of multiple sclerosis (MS) in prospective studies. Objective: To assess the time to cessation of work and receiving disability pension in MS, and how it may depend on gender, type of work and age and symptom at onset.

Method: A total of 2240 Danes with onset of definite/probable MS 1980‐1989, identified from the Danish MS‐Registry, were included. Information on social endpoints was retrieved from Statistics Denmark. Cox regression analyses were used with onset as starting point.

Results: Afferent onset symptoms [hazard ratio (HR 0.57)] and non‐physical type of work (HR 0.70) were favourable prognostic factors compared with high age at onset, physical work and efferent symptoms at onset. The mean time to disability pension was 13 years for patients with afferent/brainstem onset symptom but 8.7 years for those with efferent onset symptoms (P < 0.0001). The effect of onset symptom was reduced and the effect of sex became significant when all

covariates and age at onset were included in multivariate Cox regression.

Conclusions: Onset age, type of onset symptom and work are robust predictors of disability pension in MS. Disability pension proves to be a reliable milestone in estimation of the prognosis of MS.

Highest level of evidence: systematic review of level II studies, inception cohort study. Here we have: level III = retrospectively assembled cohort


Question V

You want to find out what to give a newlydiagnosed rheumatoid arthritis patient. One DMARD? Two products…

= question for guidelines



Outline of WS

• Introduction: EBP, RRR, ARR, NNT…• Illustrated examples• Review of evidence in small

groups• Feedback to the audience• Tips & Tricks for “quick” appraisal• Summary & take home messages

J Gen Intern Med. 2007;22:107-14.

Eur J Vasc Endovasc Surg. 2006;31:187-99.

4 groups to evaluate and appraise both reviews


Critical appraisal of secundary research

1. Is the research based upon a clearly focused question?

2. Does the research provide you the best evidence?

3. Critical appraisal of the review content itself?

4. Synthesis of the results?

Philbrick et al. Aryal et al.

1.PICO See Objectives: “evaluate the efficacyof preventivetreatments”

Mesh terms identical as ourPICO

2.EVIDENCEIn/exclusion crit

Minimal researchPublication bias

-discussed

-… reference lists…-not mentionned

-OK+discussion of excludedtrials (case reports…)-… English only…-not mentionned

3.CRITICALAPPRAISAL

Assessors?Study quality?

Clear picture onquality and LIMITSof studies

-2, consensus-In appendix

-Table 1-3 “standardsmet…”

-2?-Scottish intercollegiateGuidelines-A lot of discussion in text, nooverview in form of a table

4.SYNTHESISTable

Number of subjects perstudy not mentionnedNo Forest plot for RCT

Tables providedNo Forest plots


Philbrick et al. Aryal et al.

4.SYNTHESISPooled measure?Heterogeneity

OR + CIEfforts have been done(LOG transformation); no tests

OR + CIMentionning heterogeneity; notests

5.Interestingfindings

-Same RCTs are found by both review groups-Wide range of literature on the subject (especially Aryal)-Aryal et al: good for specialist on the field, less easy forquick appraisal-8/9 of the RCT’s from the same authors? Suspicious

6.ContentPREVENT DVT WITH STOCKINGS?

The exact incidence is difficult to estimate since the flightpopulation is healthier than the overall population.F.i. 5% per long haul flight for high risk population.-Evidence on risk factors of the patients? Small evidence-Minimum duration of flight: cut off 6 hours-Age not determined YES in patients with risk factors: level 1 (cfr. RCT) Type: ankle pressure 14-30 mmHG (Class 1 – 2)

This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2010, Issue 1


40 pages


Compression stockings applying gentle pressure to the ankle in particular, help the blood flow. This combined with leg movement helps the blood “move”.

Very large reduction in symptomless DVT

Less discomfort and swelling in the legs

Data from 9 trials, > 2800 patients.

Critical appraisal of primary research

What is the PICO of the study? And does it correspond with own PICO?

How well was the study done?

What do the results mean and could they have been due to chance?

What does this mean for my research question for my patient? Is it applicable?


What outcome do you search for?

• Patient Oriented Outcomese.g. reduction of heart attacks,

number of diabetic foot ulcers…• Disease Oriented Outcomese.g. reduction blood pressure,

improvement of HbA1C…

Tips and Tricks for primary research: n=crucial

• Number of patients, timeframe, multicentre study & number of patients per centre …

• Irrealistic inclusion criteria… “hypothetic patients” far from real life….

• Intention to treat versus Per Protocol• NNT= 1/ARR• Read the comments or editorials related to

the individual trial you focus on, cfr. BMJ


Diagnosis and management of hypocalcaemiaBMJ 2008;336;1298-1302

Plus

Unanswered questions


8 pages


2 page “critical appraisal”

• Reviews previous studies

• Summarizes the results of this study with its limitations and putting them in its clinical context

• Advises the “user”

Outline of WS



How to manage?

• Reflect on your practice on a regular basis

• Inquire, don’t advocate• Feel good about not knowing everything• Learn to ask a focused clinical question• Let someone else do the heavy lifting

Summary & Take Home Messages


– Selecting the optimal resource within the given setting requires insight in available sources, skills and practice.

– Applying findings and evidence to clinical setting might differ in different parts of the world as the findings might differ.

– These differences might be influenced by the patient’s origin as well as the practitioner’s origin.

– Be pro-active by subscribing to table of contents from major journals, providers such as Medscape, MDLink and other …. Free of charge.


• Getting acquainted with “literature” resources

• Using internet effectively: do not be scared to search; 5 minutes can/must be sufficient

• Creating easy access to favorite resources

• Learning basic commands and shortcuts•

• And …. Practice, Practice, Practice


Clinical questions, LLL and EBMHealthcare providers will learn best when learning1) is in the context of patient care2) answers their questions3) is directly applicable to their work4) does not take too much time

Information at POC: Answering clinical Qs. Ebell M; 1999

EBM = Evidence Based Medicine, LLL= Live Long Learning



We picked some examples to work out from the…

Evidence-based practice workbook, Second ed.

Paul Glasziou et al.2008

BMJ/BooksISBN 978-1-4051-6728-4

Evidence Based Practice: finding the best evidence

in 5 minutes

ESCP Lyon21-23 October 2010

Evidence Based Practice: finding the best evidence in 5 ...€¦ · finding the best evidence in 5 minutes Yolande Hanssens SIG Leader Medicine Information Clinical Pharmacy Coordinator

Documents